Kepler DR25 recovery near 372.5 days differs by -0.6 percentage points from nearby periods in 1,436 matched pairs at expected MES 6-20

We test whether recovery of added transit signals has a local loss near the 372.5-day Kepler spacecraft year. We use public DR25 INJ1 pixel-level injections and the original DR25 stellar table. The sample has temperature 4000 to 7000 K, log g > 4, expected multiple event statistic (MES) 6 to 20, duration 1.5 to less than 15 hours, and recorded transit-count field 3 to less than 5. The near-year window is 362.5 to 382.5 days. Comparison periods are 312.5 to less than 352.5 days and greater than 392.5 to 432.5 days. Unique pairs have the same sky group and similar values of nine input controls. Matching does not use recovery labels. After three pairs with an undefined label are removed, 1436 pairs remain. Recovery is 807/1436 (56.2%) near the spacecraft year and 815/1436 (56.8%) for comparison signals. The paired difference is -0.56 percentage points. Its pair bootstrap 95% interval is -3.41 to +2.30 points. Sky-group and four-channel-block intervals also include zero. All specified checks include zero; lower- and upper-period comparisons have opposite signs. The result gives no clear local loss in this matched sample. It permits modest differences and does not measure catalog reliability, planet occurrence, or the total period effect. We preserve source responses, hashes, selection and pair data, code, figures, an independent parser audit, and an exact offline repeat.

Kepler · DR25 · transit injection · search completeness · spacecraft year · matched pairs · reproducible astronomy · null result